Tamal

Associate researcher @ LCS2 | IIT Delhi |
Agentic Reasoning | Inverse Problems | Diffusion ...

[Tanmoy Chakraborty @Tanmoy_Chak]

I am delighted to announce the acceptance of 7 papers to EMNLP 2026 (Main: 4 and Findings: 3).

Congratulations to the entire lab for the hard work!!

The preprints of most of the papers are already available. The remaining will be uploaded soon.
@lcs2lab @YardiScAI @iitdelhi @emnlpmeeting #EMNLP2026

[Tanmoy Chakraborty @Tanmoy_Chak]

** 𝐀𝐝𝐚𝐩𝐭𝐢𝐯𝐞 𝐀𝐠𝐞𝐧𝐭 𝐂𝐨𝐨𝐫𝐝𝐢𝐧𝐚𝐭𝐢𝐨𝐧 **
Multi-agent systems are powerful, but they can drastically multiply inference costs. Many existing systems rely on fixed or densely activated agent pipelines without adapting computation to each query: Which agents actually need to be consulted? How deep should the reasoning go? And when is communication worth its compute cost?

Presenting 𝐆𝐑𝐀𝐃𝐄 -- Gated Routing and Adaptive Depth for Efficient Reasoning

🔗 Preprint: https://t.co/C9lHUvkPCw

GRADE optimises multi-agent reasoning by:
🧠 𝐋𝐞𝐚𝐫𝐧𝐞𝐝 𝐂𝐨𝐨𝐫𝐝𝐢𝐧𝐚𝐭𝐢𝐨𝐧: We built a hierarchical system governed by lightweight gates that jointly manage agent selection, routing depth, communication, and pruning dynamically per query.
⚙️ 𝐂𝐨𝐆𝐑𝐏𝐎 𝐓𝐫𝐚𝐢𝐧𝐢𝐧𝐠: We adapt GRPO for collaborative settings with a novel, critic-free RL recipe that assigns a shared advantage signal to all participating agents and gates during a rollout.
🔄 𝐇𝐨𝐭-𝐒𝐰𝐚𝐩𝐩𝐚𝐛𝐥𝐞 𝐄𝐱𝐩𝐞𝐫𝐭𝐬: GRADE features an Expert Registry with per-agent calibration maps. You can swap out expert models at inference time using just 64 anchor queries, without retraining the gates.

🏆 At ~17B average active parameters, GRADE outperforms all baselines on GSM8K, GPQA, and MMLUPro -- beating the strongest baseline on MMLUPro by 4.8 points while using ~39% fewer active parameters.

w/ @ScientificGhosh

Do check out many more exciting works on small models and agentic coordination being developed as part of our mega project -- 𝐏𝐚𝐫𝐚𝐦𝐚𝐧𝐮
https://t.co/Hi96YYqyzm
@lcs2lab @iitdelhi

#LLMEfficiency #MultiagentRounting #AgenticAI

[ayan sengupta @ayans007]

🔬 Parmanu (Hindi for Atom) is live.

Parmanu is part of the Computational Social Systems (LCS2) @lcs2lab at IIT Delhi, led by Prof. Tanmoy Chakraborty @Tanmoy_Chak , and is our dedicated home for Efficient Large Language Models (LLMs) and Small Language Models (SLMs).

We’re at a turning point in AI. The future won’t be defined by scaling alone - it will be shaped by efficiency, accessibility, and real-world deployability. Parmanu is our effort to push this efficiency-first vision forward. ✨🤖

🔗 Explore the project page: https://t.co/siVwaKY1GN

Why Parmanu matters 🔥

• 📚 A centralized hub for our research, with papers accepted at ICLR, ICML, NeurIPS, TACL, ACL, and TMLR
• 🛠️ Open access to tools, code, and artifacts spanning model compression, KV efficiency, PEFT, inference optimization, knowledge distillation, and model coordination
• 🧠 A growing ecosystem focused on making strong language models smarter per parameter, not just larger

What this means for the community

For researchers 👩‍🔬👨‍🔬

A curated, evolving resource tied to top-tier venues

Reproducible artifacts and principled problem formulations

A shared space to advance efficiency-centric LLM research

For practitioners 👩‍💻👨‍💻

Practical techniques to deploy LLMs under tight latency and memory budgets

Faster paths from paper → production

Tools that actually work under real deployment constraints

What’s coming in 2026 🚀🔮

• 📊 Efficient LLM/SLM leaderboards
• 🧪 Open-sourced efficient LLM artifacts
• ⚙️ More tools for compression, distillation, and inference
• 🤝 Deep integration with Hugging Face and other popular libraries

If you’re excited about efficient, sustainable, and scalable AI, check out Parmanu, share feedback, and collaborate with us. The next wave of LLMs won’t just be bigger - they’ll be leaner, faster, and more impactful. 🌟

#EfficientLLMs #SLMs #ModelCompression #InferenceOptimization #KnowledgeDistillation #AIResearch #NLP #ICLR #ICML #NeurIPS #ACL #TACL #TMLR #IITDelhi #LCS2 #Parmanu

[Unsloth AI @UnslothAI]

You can now fine-tune LLMs and deploy them directly on your phone! 🚀

We collabed with PyTorch so you can export and run your trained model 100% locally on your iOS or Android device.

Deploy Qwen3 on Pixel 8 and iPhone 15 Pro at ~40 tokens/sec.

Guide: https://t.co/8wyQLJfzeC

[Tanmoy Chakraborty @Tanmoy_Chak]

It was an honour hosting the legend @rao2z at our lab @lcs2lab, @iitdelhi yesterday. We had an engaging 2.5-hour, air-tight session discussing various aspects of LLMs. Thank you, @rao2z, for spending such valuable time with us.

[Tanmoy Chakraborty @Tanmoy_Chak]

🚀 Today I gave a keynote at ACM COMPUTE on “Rebooting NLP Teaching in the LLM Era”.

Slides: https://t.co/RITzfE4VAI

I proposed a 42-hour NLP curriculum that fuses classical statistical NLP with modern LLMs.

I'd love to hear feedback, perspectives, or experiences from the NLP community — how are you approaching NLP education today? 🙌

@Indiaacm #ACMCOMPUTE #ComputingEducation #AI4Education

[Hamel Husain @HamelHusain]

I promise you won’t be able to watch this evals meme video without laughing 🤣

Shreya and I have been working hard on the next iteration of our AI evals course. We've added some new things for students:

1. Exclusive/early access to books, tools, flashcards, and more
2. Private discord to get help anytime you're get stuck (lifetime access)
3. Live office hours + unlimited access to future cohorts so you never have to worry about timing or missing out on new material.

Reply tweet has links

[Tanmoy Chakraborty @Tanmoy_Chak]

The ICLR'26 reviewer-name leak is a reminder that we're no longer advancing science -- we're trapped in a system that often disrespects it.

Maybe it's time for AI/ML to pause and rethink the entire conference culture.

Wild idea: Stop all major AI/ML conferences for 1-2 years. Let papers live on arXiv, and be submitted to journals. Meanwhile, conferences rebuild real policies and accountability.

Will this ever happen? Probably not.
The incentives are too big. Its about MONEY and PROFIT.

But the community deserves better and we need the courage to imagine alternatives.

@iclr_conf @NeurIPSConf @RealAAAI

[Tamal @dextro_rx]

@neural_avb This one's crazy good. Loved it 🔥

[Tamal @dextro_rx]

@xdpurvesh Bruh thank you so much 🧡🧡

[Tamal @dextro_rx]

@MuskanRaina97 Beautiful 🧡🔥

[Tamal @dextro_rx]

@Fearlessproduc7 @rikognition Congratulations man 🧡🧡

[Tamal @dextro_rx]

@abg_3d @opensea @jf_nft @reverse_visuals @xdpurvesh @NihalGilbert @Vishhman @TomaszBolek @TommyBui_0110 @thepriestessnft @meerkat_space Thank you so much mahn . 🧡🌌

[ʰᵘˡᵉᵉᵇ @huleeb]

HEAT 🔥

[Swapnonil Munshi @faajil_]

"Beauty of Pink shades"

[Tamal @dextro_rx]

@thepriestessnft @opensea @jf_nft @reverse_visuals @xdpurvesh @NihalGilbert @Vishhman @TomaszBolek @TommyBui_0110 @meerkat_space Thank you 🙏🧡

[Tamal @dextro_rx]

@NihalGilbert @opensea @jf_nft @reverse_visuals @xdpurvesh @Vishhman @TomaszBolek @TommyBui_0110 @thepriestessnft @meerkat_space Thank you bruh🧡🧡

[Tamal @dextro_rx]

Outstanding mahn 🧡

[imit0r @imit0r]

SURREAL SERIES is live on @opensea! A collection of dark-themed renders that tell of a journey, with no beginning and no end. Check it out here:

https://t.co/neGcSCuT2M

[Tamal @dextro_rx]

@TommyBui_0110 @opensea @jf_nft @reverse_visuals @xdpurvesh @NihalGilbert @Vishhman @TomaszBolek @thepriestessnft @meerkat_space Thank you so much mahn 🙏🧡

[Tamal @dextro_rx]

Working on something new .
Dropping next week . Stay tuned . Peace .